本文へ移動
cccskills
無料GitHub で公開

role-model

Use when Pi should route model requests through a locally running Role-Model runtime, inspect Role-Model aliases, or diagnose Role-Model provider setup.

日本語の概要は準備中です。原文の説明を表示しています。

try-works/role-modelスター 117(GitHub 117 + cccskills 0)

実行スコア 未検証

インストール方法を見る

含まれるファイル(1)

  • SKILL.md4.5 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Role-Model For Pi

Use this skill when the user wants Pi to use Role-Model as the routing provider.

Role-Model is the routing authority. Pi sends requests to the role-model provider, and the external Role-Model runtime decides which endpoint or alias should handle the request. The Pi package only discovers the runtime, registers provider models, exposes diagnostics, and helps choose aliases.

Use slash commands only from an interactive Pi session. pi -p "/role-model status" is unsupported because Pi print mode does not currently invoke package slash commands; treat that as an upstream Pi limitation, not as proof that Role-Model routing is broken.

First check /role-model status. If setup has not completed, run /role-model setup, then use /role-model doctor to verify the endpoint, runtime version, auth state, endpoint trust, provider state, downstream discovery contract, and the current canonical provider-relative alias guidance.

For explicit provider prompts, use the provider-relative Role-Model id that Pi lists for provider role-model, for example pi --no-session --provider role-model --model baseline.remote-only -p "<prompt>". The provider-relative alias is canonical. role-model/<alias> is compatibility-only for Pi surfaces that explicitly require a qualified id.

For model selection, use /role-model alias list to inspect available aliases, /role-model alias recommended to show the runtime recommendation, and /role-model alias use <alias> or /role-model alias choose <alias> to select the alias Pi should use. Use /role-model alias refresh after the Role-Model runtime catalog changes. If someone tries a foreign id such as gpt-4o under provider role-model, send them back to /role-model alias list and /role-model alias recommended.

For runtime-owned request diagnostics, use /role-model requests to inspect recent runtime requests and /role-model explain <request-id|latest> to read the runtime's structured request detail and router decision detail for that request. Use those commands when the user asks why Role-Model chose a route, what reason codes were recorded, or whether benchmark/telemetry advisory signals showed up in the runtime diagnostics.

Aliases are Role-Model routing entries; direct models are concrete model IDs exposed by the runtime. Prefer aliases when the user wants Role-Model to route intelligently.

For taxonomy discovery, prefer the live Role-Model runtime taxonomy manifest and compact role/task endpoints when available. If the runtime is unavailable or reports an incompatible taxonomy version, fall back to the package compact taxonomy snapshot and report the version mismatch clearly.

Use progressive classification before sending a request through Role-Model: choose likely groups first, inspect likely roles next, and load task detail only for the likely role or an ambiguous role pair. Include role_model.intent metadata when the package can classify the request: role, task, preferred capabilities, modalities, tool classes, confidence, source, evidence, taxonomy version, and classification contract version. Treat low-confidence heuristic classifications as advisory unless the user gave an explicit instruction.

Role-Model benchmarks, routing decisions, endpoint eligibility, fallback, and telemetry belong to Role-Model, not to this Pi package. Use Role-Model runtime docs and UI for benchmark or routing analysis. Taxonomy-aware benchmarks and telemetry are later Role-Model phases, and this package may surface runtime-owned request diagnostics and reason codes, but it must not claim that it scores models or aggregates production telemetry itself.

If the user explicitly asks to install or launch the external Role-Model router runtime, point them to the Role-Model repository README for those instructions and make it clear that this is external runtime setup, not a side effect of installing the Pi package.

Security boundaries: remote endpoints require explicit trust, runtimes reporting authentication.required must fail closed unless an explicit supported token source exists, and the package must not manage the runtime lifecycle. Do not read, print, copy, or sync Pi auth files.

Raw HTTP calls to the runtime are debug-only fallback tools when diagnosing Pi or runtime issues. They are not the primary supported workflow when the Pi provider path is available.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

コメント

ログインなしで書き込めます。使ってみた感想、質問、動かなかった環境などを気軽にどうぞ。

まだコメントはありません。最初のコメントを書いてみませんか。

送信するとすぐに公開されます。メールアドレスや電話番号などの個人情報は書かないでください。URL は 2 つまでです。

0 / 1000 文字

このフォームは reCAPTCHA で保護されており、Google のプライバシーポリシーと利用規約が適用されます。

同じリポジトリのスキル

Build AI agents on Cloudflare Workers using the Agents SDK. Load when creating stateful agents, durable workflows, real-time WebSocket apps, scheduled tasks, MCP servers, chat applications, voice agents, or browser automation. Covers Agent class, state management, callable RPC, Workflows, durable execution, queues, retries, observability, and React hooks. Biases towards retrieval from Cloudflare docs over pre-trained knowledge.

日本語の概要は準備中です。原文の説明を表示しています。

try-works/role-model1172026年10月25日 更新

Comprehensive Cloudflare platform skill covering Workers, Pages, storage (KV, D1, R2), AI (Workers AI, Vectorize, Agents SDK), feature flags (Flagship), networking (Tunnel, Spectrum), security (WAF, DDoS), and infrastructure-as-code (Terraform, Pulumi). Use for any Cloudflare development task. Biases towards retrieval from Cloudflare docs over pre-trained knowledge.

日本語の概要は準備中です。原文の説明を表示しています。

try-works/role-model1172026年10月25日 更新

Send and receive transactional emails with Cloudflare Email Service (Email Sending + Email Routing). Use when building email sending (Workers binding or REST API), email routing, Agents SDK email handling, or integrating email into any app — Workers, Node.js, Python, Go, etc. Also use for email deliverability, SPF/DKIM/DMARC, wrangler email setup, MCP email tools, or when a coding agent needs to send emails. Even for simple requests like "add email to my Worker" — this skill has critical config details.

日本語の概要は準備中です。原文の説明を表示しています。

try-works/role-model1172026年10月25日 更新

Guides Cloudflare One Zero Trust and SASE work across Access, Gateway, WARP, Tunnel, Cloudflare WAN, DLP, CASB, device posture, and identity. Use when designing, configuring, troubleshooting, or reviewing Cloudflare One deployments. Retrieval-first: use current Cloudflare docs/API schemas instead of embedded product docs.

日本語の概要は準備中です。原文の説明を表示しています。

try-works/role-model1172026年10月25日 更新

Plans migrations from Zscaler ZIA/ZPA, Palo Alto, legacy VPN, SWG, or SASE stacks to Cloudflare One. Use for migration assessments, policy mapping, rollout plans, and parity/gap analysis.

日本語の概要は準備中です。原文の説明を表示しています。

try-works/role-model1172026年10月25日 更新

Review the changed lines of a single file in a pull request for bugs, correctness, error handling, security, and maintainability, and return structured findings.

日本語の概要は準備中です。原文の説明を表示しています。

try-works/role-model1172026年10月25日 更新

try-works のスキルをすべて見る

このスキルの問題を報告する